A smart management method and system for shared dynamic lifecycle data in families

Abstract

This invention discloses an intelligent management method and system for dynamic data of family sharing lifecycle, relating to the field of artificial intelligence technology. The method includes: allocating management permissions based on a set of benefits; constructing an initial family relationship graph; inviting family members based on the initial family relationship graph; predicting the probability of joining through a T-S fuzzy neural network; updating the initial family relationship graph; distributing a set of benefits; and generating a distribution record. Furthermore, by combining Wiener filtering and a CNN model to process speech data and generate emotional weights, dynamic benefit allocation based on the emotional state of family members is achieved, improving the personalization and accuracy of resource allocation. The method also dynamically updates the joining tendency of family members through a T-S fuzzy neural network and the family relationship graph, optimizing family sharing permission management and priority ranking, and enhancing the system's adaptability and interaction efficiency.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to an intelligent management method and system for shared dynamic lifecycle data of families. Background Technology

[0002] With the rapid development of smart home and IoT technologies, home data sharing and management has become an important research area for improving the efficiency of family member interaction and quality of life. Home data management methods based on cloud computing, big data analysis, and artificial intelligence have been widely applied, covering multiple aspects such as resource allocation, access control, and emotional interaction analysis among family members. Existing technologies collect behavioral data of family members through API interfaces and combine them with simple rule engines or statistical models to achieve resource allocation and access control. Advances in speech recognition and emotion analysis technologies have made it possible to extract emotional states from speech data. Common methods include fundamental frequency analysis, Mel-frequency cepstral coefficient extraction, and convolutional neural networks for emotion classification. These technologies provide basic support for dynamic interactions among family members and promote the intelligent development of home sharing systems. The introduction of distributed computing and edge computing technologies has significantly improved data processing efficiency and real-time performance, providing technical support for dynamic data management in home scenarios.

[0003] However, existing technologies still have shortcomings. Existing systems are unable to dynamically adjust the rights and interests allocation strategy based on the emotional state of family members, which limits the personalization and intelligence of resource utilization. When dealing with the dynamic addition of family members and permission management, existing technologies lack comprehensive modeling of complex family relationships and interaction history, making it difficult to achieve priority ranking and tendency prediction, thus affecting sharing efficiency. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides an intelligent management method and system for dynamic data of family sharing lifecycle, which solves the problem that existing systems are unable to dynamically adjust the rights and interests allocation strategy according to the emotional state of family members, thus limiting the personalization and intelligence of resource utilization. Existing technologies lack comprehensive modeling of complex family relationships and interaction history when dealing with the dynamic addition of family members and permission management, making it difficult to achieve priority ranking and tendency prediction, thereby affecting the efficiency of sharing.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, the present invention provides an intelligent management method for shared dynamic lifecycle data of families, comprising,

[0008] Collect basic rights parameters and voice data of the main account, generate voice feature vectors, input the voice feature vectors into the CNN model, output the probability vector of the emotional state of the main account, obtain the emotional weights, calculate the family emotional synergy weights, generate rights definition data for the types of rights in the basic rights parameters, use a distributed hash algorithm to generate batch numbers and rights numbers, and merge them to form a rights set;

[0009] Based on the allocation and management authority of the rights set, an initial family relationship diagram is constructed. Family members are invited based on the initial family relationship diagram. The probability of joining is predicted by the TS fuzzy neural network. The initial family relationship diagram is updated, and the rights set is distributed and the distribution record is generated.

[0010] Obtain time-series data on family members' usage, fill in missing values ​​to obtain the filled time series, input it into the LSTM model to generate hidden states, combine it with MLP to adjust and generate predicted usage frequency and segment-level representation vectors, combine it with family emotional synergy weights to calculate grey relational degree, obtain allocation weights, generate allocation records, calculate initial value balances, family members upload sharing requests, map request time points to time periods, calculate sharing priority, generate a full life cycle report of rights and interests, and store it in the local database.

[0011] As a preferred embodiment of the intelligent management method for dynamic data of family shared lifecycles as described in this invention, the step of updating the initial family relationship diagram and distributing rights sets to generate distribution records includes:

[0012] The main account sets the upper limit of family members and initial roles through the API interface, assigns management permissions to the initial roles through the rights set, constructs the initial family relationship graph, invites family members through the API interface, collects voice data of family members through the API interface, calculates the family emotional synergy weight of family members, combines the rights set and the family relationship graph, predicts the probability of joining through the TS fuzzy neural network, sorts the probability of joining in descending order, selects the family members with the highest value for priority invitation, updates the initial family relationship graph, obtains the family relationship graph, distributes the rights set based on the family relationship graph, and generates distribution records.

[0013] As a preferred embodiment of the intelligent management method for dynamic data of shared family lifecycles as described in this invention, the input LSTM model generates hidden states, and the MLP is combined to adjust and generate prediction usage frequency and segment-level representation vectors, including:

[0014] The usage time series data of family members is obtained through the API interface, normalized to obtain a normalized time series, missing values ​​are imputed in the normalized time series to obtain an imputed time series, which is then input into the LSTM model to generate hidden states. The model is then adjusted using MLP to generate predicted usage frequency and segment-level representation vectors.

[0015] As a preferred embodiment of the intelligent management method for dynamic data of family shared lifecycles as described in this invention, the step of calculating gray relational degree by combining family emotional synergy weights, obtaining allocation weights, and generating allocation records includes:

[0016] Calculate the grey relational degree, combine the predicted usage frequency and segment-level representation vector to obtain the allocation weights, allocate the initial values, and generate allocation records.

[0017] As a preferred embodiment of the intelligent management method for dynamic data of family shared lifecycle as described in this invention, the step of generating a full lifecycle report of rights and interests and storing it in a local database includes:

[0018] The allocation record is uploaded to the edge node via the API interface, the initial value balance is calculated, and family members upload sharing requests via the API interface. The request time point is mapped to the corresponding time period, the sharing priority of the requesting family member in the corresponding time period is calculated, the initial value balance is allocated, and the sharing record is generated.

[0019] The rights set, issuance records, usage records, and status markers are arranged vertically to generate a rights lifecycle report, which is then stored in a local database.

[0020] As a preferred embodiment of the intelligent management method for dynamic data of shared family lifecycles described in this invention, the method involves: collecting basic rights parameters and voice data of the main account through an API interface to generate a voice feature vector, including:

[0021] The system collects basic parameters of the main account's rights and voice data through the API interface, and uses the Wiener filtering algorithm to filter the voice data to generate voice feature vectors.

[0022] As a preferred embodiment of the intelligent management method for dynamic data of shared family lifecycles described in this invention, the method includes: using a distributed hash algorithm to generate batch numbers and rights numbers, merging them to form a rights set, including:

[0023] The speech feature vector is converted into a time-spectrum graph through short-time Fourier transform, and then converted into a Mel spectrum graph through Mel filter. The graph is then input into a CNN model, and the output is the probability vector of the emotional state of the main account. The family emotional synergy weight is obtained, and the recommendation rule base is queried to adjust the types of rights and interests, generating rights and interests definition data.

[0024] A distributed hash algorithm is used to generate batch numbers and equity numbers for the equity definition data, including batch numbers and serial numbers. The equity numbers and batch numbers are then incorporated into the equity definition data to generate an equity set.

[0025] Secondly, this invention provides an intelligent management system for shared dynamic lifecycle data within a family, comprising:

[0026] The data acquisition and preprocessing module is used to collect the main account's rights parameters and voice data through the API interface, and to perform Wiener filtering on the voice data.

[0027] The sentiment analysis and feature extraction module is used to generate feature vectors based on speech data, analyze emotional states using a CNN model, and calculate family emotional synergy weights.

[0028] The rights definition and number generation module is used to adjust rights parameters based on family emotional collaboration weights and recommendation rule bases, generate rights definition data, batch numbers and rights numbers, and form a rights set.

[0029] The family relationship building and invitation module is used to predict the probability of family members joining based on the initial family relationship graph and TS fuzzy neural network, update the relationship graph and distribute the rights set;

[0030] The dynamic allocation and lifecycle management module is used to calculate allocation weights using LSTM models and grey relational analysis, process sharing requests, generate full lifecycle reports, and manage equity status.

[0031] Thirdly, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, it implements any step of the intelligent management method for shared dynamic lifecycle data of a family as described in the first aspect of the present invention.

[0032] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the intelligent management method for shared dynamic lifecycle data of a family as described in the first aspect of the present invention.

[0033] The beneficial effects of this invention are as follows: This invention constructs an initial family relationship diagram by allocating management permissions based on a set of rights, invites family members based on the initial family relationship diagram, predicts the probability of joining tendency through a TS fuzzy neural network, updates the initial family relationship diagram, distributes a set of rights, and generates a distribution record; it realizes dynamic rights allocation based on the emotional state of family members, improves the personalization and accuracy of resource allocation, optimizes family shared permission management and priority ranking, and enhances the adaptability and interaction efficiency of the system. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a flowchart of the intelligent management method for shared dynamic lifecycle data of families in Example 1.

[0036] Figure 2 This is a schematic diagram of the intelligent management system for sharing dynamic lifecycle data of families in Example 1. Detailed Implementation

[0037] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0038] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0039] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0040] Example 1, referring to Figure 1 and Figure 2 This is the first embodiment of the present invention, which provides an intelligent management method for shared dynamic lifecycle data of families, including the following steps:

[0041] S1. Collect basic rights parameters and voice data of the main account through API interface, generate voice feature vector, input the voice feature vector into CNN model, output the probability vector of the emotional state of the main account, obtain the emotional weight, calculate the family emotional coordination weight, generate rights definition data for the types of rights in the basic rights parameters, use distributed hash algorithm to generate batch number and rights number, and merge to form rights set;

[0042] Specifically, basic parameters of the main account's rights and voice data are collected through the API interface to generate voice feature vectors, including:

[0043] The system collects basic parameters of the main account's rights and voice data through the API interface, and then uses the Wiener filtering algorithm to filter the voice data.

[0044] The basic parameters of the rights include the type of rights (e.g., music membership and coupons), the unit of measurement (e.g., number of times and amount), the initial value (e.g., 30 plays + 100 yuan of data), the validity period type (e.g., fixed period or dynamic validity period after activation), and the sharing rules (e.g., allocation based on emotional weight, role, and usage frequency).

[0045] Based on speech data, speech feature vectors are generated, including calculating the fundamental frequency using the autocorrelation method, calculating the amplitude through short-time energy analysis, extracting MFCC coefficients using Mel frequency cepstral analysis, and calculating the speech rate using a syllable detection algorithm. The results are then stacked horizontally.

[0046] This invention employs the Wiener filtering algorithm to denoise speech data and combines it with autocorrelation, short-time energy analysis, Mel frequency cepstral analysis, and syllable detection algorithms to generate speech feature vectors. This enables multi-dimensional feature extraction of the main account's speech data, significantly improving the accuracy and robustness of sentiment analysis, constructing richer speech feature vectors, effectively reducing the impact of noise interference on sentiment recognition, and providing high-quality input data for subsequent CNN-based sentiment state classification. This further improves the accuracy of calculating family member sentiment weights, laying the foundation for dynamic rights allocation and enabling the system to recommend personalized resources based on the real-time sentiment states of family members.

[0047] Furthermore, a distributed hash algorithm is used to generate batch numbers and equity numbers, which are then merged to form an equity set, including:

[0048] Construct a CNN model, including an input layer, convolutional layers, activation function layers, pooling layers, flattening layers, fully connected layers, and an output layer;

[0049] The CNN model was trained using the RAVDESS dataset.

[0050] The speech feature vector is converted into a time-spectrum graph (a two-dimensional matrix including time and frequency axes) through short-time Fourier transform. Then, it is converted into a Mel spectrum graph using a Mel filter. This graph is input into a CNN model and outputs a probability vector of the emotional state of the main account, such as happiness = 0.8 and fatigue = 0.4. The probabilities are sorted in descending order to select the emotional state corresponding to the maximum value. The corresponding probability is then normalized to obtain the emotional weight.

[0051] The formula for calculating the weight of family emotional synergy is as follows:

[0052] ,

[0053] in, Weighting of family emotional coordination For the first Weighting coefficients for each family member, For the first The emotional weight of each family member For the first The first family member and the first The intensity of emotional interactions among family members is obtained by acquiring historical voice interaction frequencies through an API interface and then normalizing them. Number of family members;

[0054] Based on the summation of the family emotional synergy weights, the recommendation rule base is queried to adjust the types of rights. For example, if the summation of the family emotional synergy weights is greater than or equal to the threshold of the summation of the family emotional synergy weights, then entertainment is recommended.

[0055] The recommendation rule base refers to the rule base generated based on historical data analysis (such as the correlation between common family scenarios and rights types, obtained through API interfaces) and custom rules of the main account.

[0056] The adjusted types of rights, units of measurement, initial values, validity periods, and family emotional synergy weights are arranged horizontally to generate rights definition data.

[0057] A distributed hash algorithm is used to generate batch numbers for the rights definition data, including timestamps and random strings (obtained using a random number generator), and rights numbers, including batch numbers and serial numbers, are generated for the rights definition document data.

[0058] The rights number and batch number are incorporated into the rights definition data to generate a rights set.

[0059] By constructing a CNN model and training it on the RAVDESS dataset, speech feature vectors are converted into Mel spectrogram input models, outputting emotional state probability vectors. Based on a family emotional synergy weight formula, the types of rights are dynamically adjusted. Finally, a distributed hash algorithm is used to generate batch numbers and rights numbers, forming a rights set. This achieves efficient management and dynamic optimization of family shared resources. Through sentiment analysis using the CNN model and Mel spectrogram, combined with the family emotional synergy weight formula, dynamic adjustment of rights types based on emotional interaction is achieved. The introduction of the distributed hash algorithm generates unique batch numbers and rights numbers using timestamps and random strings, ensuring the traceability and security of the rights set. This not only improves data processing efficiency but also significantly enhances the personalization and scenario adaptability of resource allocation through an emotion-driven dynamic adjustment mechanism, optimizing the sharing experience for family members.

[0060] S2. Based on the allocation and management authority of the rights set, construct an initial family relationship diagram, invite family members based on the initial family relationship diagram, predict the joining tendency probability through the TS fuzzy neural network, update the initial family relationship diagram, distribute the rights set, and generate distribution records;

[0061] Specifically, the initial family relationship diagram is updated, and a set of benefits is distributed, generating distribution records, including:

[0062] The main account sets the maximum number of family members and initial roles (such as parents and children) through the API interface, and assigns management permissions to the initial roles through the rights set. For example, when the family emotional synergy weight in the rights set is happiness, the initial role is assigned high management permissions (e.g., initial role weight 1.5), and when the family emotional synergy weight in the rights set is fatigue, the initial role is assigned low management permissions (e.g., initial role weight 1.2). The rights number is set as the family ID, and the initial family relationship graph is constructed.

[0063] The initial family relationship graph defines family members (including initial roles, family emotional synergy weights, and role weights) as nodes, relationships between members as edges, and the intensity of emotional interaction as edge weights.

[0064] Construct a TS fuzzy neural network, including an input layer, a fuzzification layer, a fuzzy rule layer, and an output layer;

[0065] Fuzzy rules are set based on business logic. For example, if the role is a parent and the emotional interaction intensity is >0.8, then the tendency probability is 0.9. The TS fuzzy neural network is then trained.

[0066] Based on the initial family relationship diagram, family members are invited through the API interface, and voice data of family members are collected through the API interface. The family emotional synergy weight of family members is calculated. Combining the rights set and the family relationship diagram, the probability of joining is predicted through the TS fuzzy neural network. The joining probability is sorted in descending order, and the family members with the highest value are selected for priority invitation. The initial family relationship diagram is updated to obtain the family relationship diagram.

[0067] Updating the initial family relationship graph refers to adding new nodes to it, including the roles of family members, the weight of family emotional synergy, and the probability of joining.

[0068] The distribution of rights sets based on family relationship diagrams involves sorting the nodes in the diagram in descending order of their probability of joining, selecting the family members with the highest probability to distribute the rights sets, marking the distributed rights sets as pending activation, and generating distribution records, including the rights card number, target member ID, and validity period.

[0069] By constructing an initial family relationship graph and combining it with a TS fuzzy neural network to predict the probability of family members joining, and by dynamically updating and distributing the rights set based on the family relationship graph, the intelligent and dynamic management of family member permissions and resource allocation is realized. This significantly improves the system's adaptability to dynamic joining scenarios. Through the family relationship graph and the TS fuzzy neural network, dynamic priority sorting and permission allocation are realized, improving the interaction efficiency and user participation of the family sharing system. By dynamically distributing the rights set and marking the pending activation status, combined with the generation of distribution records, the entire lifecycle of rights is visualized and managed. This not only improves the transparency and traceability of resource allocation, but also adapts to dynamic changes such as family members joining or leaving by dynamically updating the family relationship graph, which is significantly better than the static or semi-static resource management methods in the existing technology.

[0070] S3. Obtain the usage time series data of family members through the API interface, fill in the missing values ​​to obtain the filled time series, input it into the LSTM model to generate hidden states, combine it with MLP to adjust and generate the predicted usage frequency and segment-level representation vector, combine it with family emotional synergy weight to calculate the grey relational degree, obtain the allocation weight, generate allocation records, calculate the initial value balance, family members upload sharing requests, map the request time point to the time period, calculate the sharing priority, generate a rights and interests life cycle report, and store it in the local database.

[0071] Specifically, the hidden states are generated by inputting the LSTM model, and then adjusted using the MLP to generate prediction frequency and segment-level representation vectors, including:

[0072] The usage time series data of family members can be obtained through the API interface, such as the number of music plays or data consumption in seven days, and then normalized to obtain a normalized time series.

[0073] Imputing missing values ​​in a normalized time series yields an imputed time series, using the following formula:

[0074] ,

[0075] ,

[0076] ,

[0077] in, For the first Family members at the target time point The initial fill value, The time window length, For the first Learnable bandwidth parameters for each family member For the first Task perception weights for each family member For the first Family members in historical time Normalized usage value, For the first Family members in historical time The normalization uses the mask of both values, indicating that if If observed, the value is 1; otherwise, it is 0. For Gaussian kernel function, For the first Family members at the target time point The fill value, For the number of family members, For the first The weight of family emotional synergy among individual family members For the first The role weight of each family member For the first Learnable bandwidth parameters for each family member and For the first Family members in historical time The normalized values ​​are used for both the mask and the normalized values.

[0078] Based on the time series data imputation, the input is to an LSTM model to generate hidden states, which are then adjusted using an MLP (both models are trained by collecting historical usage data from household users through an API interface) to generate predicted usage frequency and segment-level representation vectors, as shown in the formula:

[0079] ,

[0080] ,

[0081] ,

[0082] ,

[0083] ,

[0084] in, For time The dynamic Gaussian mixture distribution represents the probability density for predicting and filling in the time series. For LSTM models in time The hidden state, The gating function of the MLP is initialized with random values ​​and optimized through training with a variational autoencoder. The potential cluster state transition probabilities are obtained through the LSTM model. For time The latent representation is obtained based on sampling of the latent cluster state transition probabilities. It is a static Gaussian mixture fundamental distribution. The cluster mean is the value of a static Gaussian mixture distribution, representing a classic usage pattern for imputing time series data. For time Time The potential cluster state sequence is obtained by recursively processing the time series using LSTM to fill in the gaps. For the first Family members at time Predicted usage frequency It follows a normal distribution. and For time Potential representation The mean and standard deviation, For the first Each prototype feature vector represents a typical model over a time period. The initial values ​​are random vectors, obtained through soft-DTW optimization. The distance matrix is... and The prototype feature matrix and hidden state matrix are obtained by stacking the prototype feature vectors and hidden states using tensors. To be the optimal alignment matrix, The alignment matrix represents the alignment method between the time series and the prototype feature vectors. Let be the set of all possible alignment matrices, representing the time-warped path space. For the first The family member in Segment-level representation vectors for each time period, For the first The end time of each time period is obtained by extracting the time period boundaries based on the optimal alignment matrix and then normalizing them. For distance.

[0085] By acquiring time-series data on family members' usage through API interfaces, and using a Gaussian kernel function combined with task-aware weights and role weights to impute missing values, a time series with imputed values ​​is generated. This effectively compensates for the shortcomings of traditional methods when dealing with incomplete data and avoids prediction bias caused by missing data. The LSTM model captures long-term dependencies in the time series, and combined with the nonlinear adjustment capability of the MLP model, predictive usage frequency and segment-level representation vectors are generated. This fully explores the usage patterns and behavioral characteristics of family members in different time periods. By introducing dynamic Gaussian mixture distribution and soft-DTW optimized prototype feature vectors, the dynamic changes and typical patterns of the time series can be accurately represented, thus providing a more reliable data foundation for subsequent resource allocation. This effectively adapts to the actual needs of different family members and improves the intelligence level of the system.

[0086] Furthermore, by combining factors such as family emotional synergy weights to calculate grey relational degree, allocation weights are obtained, and allocation records are generated, including:

[0087] The grey relational degree is calculated by combining the predicted usage frequency and the segment-level representation vector to obtain the assigned weights, using the following formula:

[0088] ,

[0089] ,

[0090] ,

[0091] in, For gray relational degree, For the first The family member in The feature sequence for each time period is obtained by horizontally stacking segment-level representation vectors, family emotional synergy weights, role weights, predicted usage frequency, scene adaptation factors, and joining tendency probabilities. The ideal weight sequence is obtained by retrieving historical allocation records from the API interface. For the dimension index of the feature sequence, The resolution coefficient, For the first The family member in Scene adaptation factors for a given time period The total number of scenarios is preset by the actual application scenarios. For the first The weight of each scenario For the first The feature values ​​for each scene are obtained from the recommendation rule base. For the first The environmental feature values ​​for each scene are obtained by acquiring environmental data, such as light intensity, through an API interface and then normalizing it. For the first The family member in Predicted usage frequency for a given time period For the first The family member in Weighting of each time period This is a coefficient representing the role's weight. For the first The role weight of each family member To predict the coefficient of usage frequency, The coefficient representing the weighting of family emotional synergy. The coefficient of grey relational degree. The coefficient of the scene adaptation factor. For the first Probability of joining a family member coefficient, For the first The first family member and the first The intensity of emotional interaction among family members, Weighting for family emotional synergy;

[0092] Based on the allocation weight, the initial value is allocated. For example, the allocation weight is sorted in descending order, the family member corresponding to the largest allocation weight is selected, a high initial value is allocated, the activated status is marked, and an allocation record is generated, including family member ID, time period, allocation weight, initial value share, rights number, and allocation status.

[0093] By using grey relational analysis, multi-dimensional features such as family emotional synergy weight, role weight, predicted usage frequency, scenario adaptation factor, and participation tendency probability are comprehensively modeled to generate allocation weights. This dynamically reflects the emotional state, role importance, and scenario needs of family members. By introducing scenario adaptation factors and environmental feature values, allocation strategies can be adjusted according to actual application scenarios to ensure that resource allocation is highly matched with the real-time needs of family members. Grey relational analysis provides a scientific basis for allocation decisions by quantifying the correlation between multi-dimensional features and ideal allocation weights, and has greater flexibility and adaptability. It not only improves the fairness of rights and interests allocation, but also optimizes resource utilization efficiency by sorting and dynamically adjusting allocation weights in descending order, providing personalized and scenario-based management capabilities for family sharing systems.

[0094] Furthermore, a full lifecycle report of rights and interests is generated and stored in a local database, including:

[0095] The allocation record is uploaded to the edge node (such as a smart home gateway and a smart router) via the API interface, and the initial value balance is calculated as the difference between the initial value and the initial value share.

[0096] Family members can upload sharing requests via the API interface, including the requesting family member ID, target benefit number, request time, and requested share.

[0097] Map the requested time point to a corresponding time period, such as weekdays or weekends, using the following formula:

[0098] ,

[0099] ,

[0100] in, For the requested time point The corresponding time period index, For the first The central time point of each time period To request a specific time point, The number of time periods, for example, weekdays = 1;

[0101] Calculate the sharing priority of requesting family members within the corresponding time period, sort them in descending order, and select the family member with the highest sharing priority for initial value balance allocation. This includes allocating the requested share to the requesting family member if the initial value balance is sufficient, otherwise rejecting the allocation. A sharing record is generated, including the requesting family member ID, corresponding time period, rights number, allocated share, validity period, and sharing priority. The formula is:

[0102] ,

[0103] in, For the first The family member in The priority for congratulations corresponds to the time period. , , as well as For the first The family member in The prediction frequency, weighting, probability of incorporating bias, and intensity of sentiment interaction for each corresponding time period. and For the first The family member in The prediction frequency and weighting for each corresponding time period;

[0104] Based on shared records and issuance records, the dynamic validity period is calculated. If the dynamic validity period expires, the rights are marked as forfeited and pushed through the API interface.

[0105] ,

[0106] in, The validity period is dynamic. The basic validity period is derived from the basic parameters of the rights and interests. The sentiment adjustment factor indicates the degree to which sentiment adjusts for validity period.

[0107] The rights set, issuance records, usage records, and status markers (including pending activation, activated, and abandoned) are arranged vertically to generate a rights lifecycle report, which is then stored in a local database.

[0108] By mapping the time points of sharing requests to corresponding time periods and calculating sharing priorities based on predicted usage frequency, allocation weights, joining probability, and emotional interaction intensity, dynamic priority ranking and resource allocation for family members' sharing requests are achieved. This ensures the fairness and efficiency of resource allocation and overcomes the inefficiency of static allocation or manual intervention in traditional systems. The calculation of dynamic validity period, combined with emotional adjustment factors, allows the validity period of rights to be adaptively adjusted according to the emotional state of family members. This is more flexible than the traditional method with fixed validity period and extends the efficiency of resource utilization. By vertically arranging the rights set, issuance records, usage records, and status markers to generate a full lifecycle report and storing it in a local database, traceable records of the entire rights management process are achieved, providing a reliable guarantee for the long-term stable operation of the family sharing system.

[0109] This embodiment also provides an intelligent management system for sharing dynamic lifecycle data within a family, including:

[0110] The data acquisition and preprocessing module is used to collect the main account's rights parameters and voice data through the API interface, and to perform Wiener filtering on the voice data.

[0111] The sentiment analysis and feature extraction module is used to generate feature vectors based on speech data, analyze emotional states using a CNN model, and calculate family emotional synergy weights.

[0112] The rights definition and number generation module is used to adjust rights parameters based on family emotional collaboration weights and recommendation rule bases, generate rights definition data, batch numbers and rights numbers, and form a rights set.

[0113] The family relationship building and invitation module is used to predict the probability of family members joining based on the initial family relationship graph and TS fuzzy neural network, update the relationship graph and distribute the rights set;

[0114] The dynamic allocation and lifecycle management module is used to calculate allocation weights using LSTM models and grey relational analysis, process sharing requests, generate full lifecycle reports, and manage equity status.

[0115] This embodiment also provides a computer device applicable to the intelligent management method for shared dynamic lifecycle data of families, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize the intelligent management method for shared dynamic lifecycle data of families as proposed in the above embodiment.

[0116] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0117] This embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the intelligent management method for dynamic data of a shared lifecycle for families as proposed in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0118] In summary, this invention constructs an initial family relationship graph by allocating management permissions based on a set of rights, invites family members based on the initial family relationship graph, predicts the probability of joining through a TS fuzzy neural network, updates the initial family relationship graph, distributes a set of rights, and generates a distribution record; it realizes dynamic rights allocation based on the emotional state of family members, improves the personalization and accuracy of resource allocation, optimizes family shared permission management and priority ranking, and enhances the system's adaptability and interaction efficiency.

[0119] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent management of home-oriented shared lifecycle dynamic data, characterized in that: Comprise, Collecting the basic parameters of the rights and interests and voice data of the primary account, generating a voice feature vector, inputting the voice feature vector into a CNN model, outputting a probability vector of the emotional state of the primary account, sorting the probability vector in descending order, screening the emotional state corresponding to the maximum value, normalizing the probability vector corresponding to the emotional state to obtain an emotional weight, calculating a family emotional synergy weight, the formula being: ， wherein, is a family emotional coordination weight, is a weight coefficient of the th family member, is an emotional weight of the th family member, is an emotional interaction intensity between the th family member and the th family member, which is obtained by normalizing the historical voice interaction frequency through an API interface, is the number of family members; The basic parameters of the rights and interests include the type of rights and interests, the unit of measurement, the initial value, the type of validity period and the sharing rule; Based on the sum average of the family emotional synergy weight, querying the recommendation rule library, adjusting the type of rights and interests in the basic parameters of the rights and interests, generating rights and interests definition data, using a distributed hash algorithm to generate a batch number and a rights and interests number for the rights and interests definition data, and merging to form a rights and interests set; Based on the rights and interests set, the management authority is distributed, the initial family relationship graph is constructed, the family members are invited based on the initial family relationship graph, the joining tendency probability is predicted through a T-S fuzzy neural network, the initial family relationship graph is updated, and the rights and interests set is distributed to generate a distribution record; Obtaining the use time series data of the family members, filling in the missing values to obtain a filled time series, inputting the LSTM model to generate a hidden state, adjusting the hidden state using an MLP to generate a predicted use frequency and a segment-level feature vector, the formula being: ， ， ， ， ， in, For time The dynamic Gaussian mixture distribution represents the probability density for predicting and filling in the time series. For LSTM models in time The hidden state, The gating function of the MLP is initialized with random values ​​and optimized through training with a variational autoencoder. The potential cluster state transition probabilities are obtained through the LSTM model. For time The latent representation is obtained based on sampling of the latent cluster state transition probabilities. It is a static Gaussian mixture fundamental distribution. The cluster mean is the value of a static Gaussian mixture distribution, representing a classic usage pattern for imputing time series data. For time Time The potential cluster state sequence is obtained by recursively processing the time series using LSTM to fill in the gaps. For the first Family members at time Predicted usage frequency It follows a normal distribution. and For time Potential representation The mean and standard deviation, For the first Each prototype feature vector represents a typical model over a time period. The initial values ​​are random vectors, obtained through soft-DTW optimization. It is a distance matrix. and The prototype feature matrix and hidden state matrix are obtained by stacking the prototype feature vectors and hidden states using tensors. To be the optimal alignment matrix, The alignment matrix represents the alignment method between the time series and the prototype feature vectors. Let be the set of all possible alignment matrices, representing the time-warped path space. For the first The family member in Segment-level representation vectors for each time period, For the first The end time of each time period is obtained by extracting the time period boundaries based on the optimal alignment matrix and then normalizing them. For distance; Based on the predicted use frequency and the segment-level feature vector, the gray correlation degree is calculated in combination with the family emotional synergy weight to obtain an allocation weight, the initial value is allocated through the allocation weight to generate an allocation record, the initial value balance is calculated, the family members upload a sharing request, the request time point is mapped to a time period, the sharing priority is calculated, the rights and interests full life cycle report is generated, and stored in the local database; The formula for calculating the gray correlation degree in combination with the family emotional synergy weight based on the predicted use frequency and the segment-level feature vector to obtain the allocation weight is: ， ， ， in, For gray relational degree, For the first The family member in The feature sequence for each time period is obtained by horizontally stacking segment-level representation vectors, family emotional synergy weights, role weights, predicted usage frequency, scene adaptation factors, and joining tendency probabilities. The ideal weight sequence is obtained by retrieving historical allocation records from the API interface. For the dimension index of the feature sequence, The resolution coefficient, For the first The family member in Scene adaptation factors for a given time period The total number of scenarios is preset by the actual application scenarios. For the first The weight of each scenario For the first The feature values ​​for each scene are obtained from the recommendation rule base. For the first The environmental feature values ​​for each scenario are obtained by acquiring environmental data through an API interface and then normalizing it. For the first The family member in Predicted usage frequency for a given time period For the first The family member in Weighting of each time period This is a coefficient representing the role's weight. For the first The role weight of each family member To predict the coefficient of usage frequency, The coefficient representing the weighting of family emotional synergy. The coefficient of grey relational degree. The coefficient of the scene adaptation factor. For the first Probability of joining a family member coefficient, For the first The first family member and the first The intensity of emotional interaction among family members, Weighting for family emotional coordination.

2. The method for smart management of home-oriented shared life cycle dynamic data according to claim 1, characterized in that: The generation of the rights and interests full life cycle report and storage in the local database includes: Uploading the allocation record to the edge node through the API interface, obtaining the initial value balance by calculating the difference between the initial value and the initial value share in the allocation record; The family members upload a sharing request through the API interface, map the request time point to the corresponding time period, calculate the sharing priority of the request family members in the corresponding time period, allocate the initial value balance, and generate a sharing record; Vertically arranging the rights and interests set, the distribution record, the use record and the state marker to generate the rights and interests full life cycle report, and storing in the local database.

3. The method for smart management of home-oriented shared life cycle dynamic data according to claim 2, characterized in that: The collection of the basic parameters of the rights and interests and voice data of the primary account, and the generation of the voice feature vector include: Collecting the basic parameters of the rights and interests and voice data of the primary account through the API interface, filtering the voice data using a Wiener filtering algorithm to generate a voice feature vector.

4. The method for smart management of home-oriented shared life cycle dynamic data according to claim 3, characterized in that: The use of a distributed hash algorithm to generate a batch number and a rights and interests number for the rights and interests definition data, and merging to form a rights and interests set includes: The voice feature vector is converted into a time-frequency spectrum graph by short-time Fourier transform, and a Mel filter is used to convert it into a Mel spectrum graph, which is input into a CNN model to output a probability vector of the emotional state of the main account, obtain a family emotional coordination weight, query a recommendation rule library to adjust the equity type, and generate equity definition data; A distributed hash algorithm is used to generate a batch number and an equity number for the equity definition data, including a batch number and a serial number, and the equity number and the batch number are integrated into the equity definition data to generate an equity set.

5. A smart management system for home sharing lifecycle dynamic data, based on the smart management method for home sharing lifecycle dynamic data according to any one of claims 1-4, characterized in that: It includes, The data acquisition and preprocessing module is used to collect the main account equity parameters and voice data through the API interface, and to perform Wiener filter processing on the voice data; The sentiment analysis and feature extraction module is used to generate a feature vector based on voice data, analyze emotional states using a CNN model, and calculate a family emotional coordination weight; The equity definition and number generation module is used to adjust equity parameters based on the family emotional coordination weight and the recommendation rule library, generate equity definition data, batch numbers, and equity numbers, and form an equity set; The family relationship construction and invitation module is used to predict the probability of family members joining based on the initial family relationship graph and the T-S fuzzy neural network, update the relationship graph, and distribute the equity set; The dynamic allocation and life cycle management module is used to calculate allocation weights using an LSTM model and a gray correlation degree, process sharing requests, generate a full life cycle report, and manage equity status. 6.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is characterized in that: The processor executes the computer program to realize the steps of the intelligent management method of dynamic data of the family sharing life cycle of any one of claims 1-4.

7. A computer readable storage medium having stored thereon a computer program, characterized in that: The computer program is executed by the processor to realize the steps of the intelligent management method of dynamic data of the family sharing life cycle of any one of claims 1-4.

Images